Background: The HLA-DRB1 shared epitope (SE) is a risk factor for the development of rheumatoid arthritis (RA) and the production of anti-citrullinated protein antibodies (ACPAs) in RA patients. Our objective was to examine the real-world effectiveness of abatacept versus tumor necrosis factor inhibitors (TNFi) in patients with RA who were SE and anti-cyclic citrullinated peptide antibody (anti-CCP3) positive.

Methods: Abatacept or TNFi initiators who were SE + and anti-CCP3+ (> 20 U/mL) at or prior to treatment and had moderate or high CDAI score (> 10) at initiation were identified. The primary outcome was mean change in CDAI score over six months. Analyses were conducted in propensity score (PS)-trimmed and -matched populations overall and a biologic-experienced subgroup. Mixed-effects models were used.

Results: In the overall PS-trimmed (abatacept, n = 170; TNFi, n = 157) and PS-matched cohorts (abatacept, n = 111; TNFi, n = 111), there were numerically greater improvements in mean change in CDAI between abatacept and TNFi but were not statistically significant. Similar trends were seen for biologic-experienced patients, except that statistical significance was reached for mean change in CDAI in the PS-trimmed cohort (abatacept, 12.22 [95% confidence interval (95%CI) 10.13 to 14.31]; TNFi, 9.28 [95%CI 7.08 to 11.48]; p = 0.045).

Conclusion: In this real world cohort, there were numerical improvements in efficacy outcomes with abatacept over TNFi in patients with RA who were SE + and ACPA+, similar to results from a clinical trial population The only statistically significant finding after adjusting for covariates was greater improvement in CDAI with abatacept versus TNFi in the bio-experienced PS-trimmed cohort..